The justice department suit against Charleston County has motivated me to air a couple of thoughts I've held for some time about the election process. I would urge the public and lawmakers to consider two changes to our voting procedures-- cumulative voting and random ballots.
First of all, let me say I don't believe any system will guarantee the intent of every voter is tallied correctly. Florida's problem relates to the use of punch card ballots, which are preserved and can be manually recounted. In Charleston County, we use machines which make that impossible. The percentage of voters in Charleston County who don't press the buttons on the voting machine to correctly register their intended choices is probably close to the percentage of voters using punch card ballots who don't punch them properly. As a poll manager in Charleston County, I have seen that many voters (about one in twenty, I would hazard as an informal estimate) push the name on the ballot, instead of the lighted button next to the name. In the last election, the machines were set up not to record a voters ballot when the green "Vote" was pressed unless at least one office or question was selected. Many voters would dilligently attempt to vote their choices (at least I assume that's what they were doing behind the closed curtain), but when they left, their vote had not been recorded. We (the poll managers) would stop them and tell them, "You have to push the green button." "I did," the voter would reply. It's hard to guess who is more frustrated. "You have to vote for at least one thing for your vote to count," we explain. "I did," the voter would insist. Checking the machine, we would see that all the lights on the ballot were still flashing-- none of the buttons had been pressed, and we would explain that to select a particular candidate or question, you have to press the flashing button next to the name, not the name itself. Once that's done, however, and the green "Vote" button is pressed, there is no opportunity to second guess the voter's intent, as with punch cards or paper ballots.
Certainly, there are other irregularities as well. Some groups complain that voters who show up at the wrong polling place are disenfranchised by poll workers who don't tell them where they're supposed to vote, and other groups claim that some poll workers allow voters to vote who don't present proper identification. (Internet voting could alleviate many such problems, but if this technology is used, voters shouldn't be allowed to vote at home. Voting should be done only on internet connected machines at the public polling place on election day.) Still, every poll manager I have worked with, regardless of race or party affiliation, regardless of depth of understanding of voting machines and election laws, are wholeheartedly dedicated to making sure that the election is conducted fairly. That fact reinforces my faith in the election process, and makes me proud to be an American.
Nevertheless, I have some suggestions that I think would make the system better. The first is cumulative voting. With it, the entire basis for the Justice Department suit against Charleston County goes away, and we do not have to resort to parochial single member districts and all the politics associated with drawing district lines. The concept is simple. There are nine members on County Council, and each voter can cast a vote for one canditate for each seat up for election. That's a total of nine votes (or however many seats are up for election) for each voter. Under a cumulative voting system, each voter is given as many votes as there are positions to be filled and allowed to cast those votes for one candidate or distribute them in any way among the candidates. Minority voters can thus concentrate their votes on candidates of their choice, thus assuring proportional representation. Cumulative voting is more prevalant in corporate shareholder elections than government elections, but there is at least one precedent, in Illinois. (Note, the web page reference I originally provided with this letter, http://www.prairienet.org/icpr/, has apparently expired. For a more recent one, available when checked on 07/12/03, see http://www.fairvote.org/cumulative/)
Since County Council seats in Charleston are not strictly at large, some modification of the simple cumulative voting scheme would be necessary. In a strictly at large system, both voting and seats are district wide. In Charleston County, voting is county wide but each candidate runs for a particular seat. It is still possible to apply cumulative voting, and the outcome would be determined the same way, with the top vote getting candidate for each seat winning. For example, if there were five seats to be filled, but one of the seats were more heavily contested, many voters county wide might cast all five of their votes in that race. The losing candidate for that seat might still have gotten more votes than the winning candidates for some of the other seats, and under a strictly at large system would have won a seat. Since candidates in Charleston County run for a particular seat, though, the highest vote getter for each seat would be the winner, regardless of the voting for the other seats.
One might object to my proposal because it could result in voters concentrating their votes on candidates from their district, thus negating the benefit of at-large voting whereby each council member is answerable to the entire county. In response, I would offer two observations. First, just because voters don't vote for candidates outside their district doesn't mean they can't. It is that possibility that forces council members to consider the interests of voters outside their districts. My second observation is that even under the present system, many voters (I estimate about 10% based on my own statistical analysis of a recent county wide election) only vote for candidates from their own district anyway. Some of these voters, it is likely, abstain intentionally, perhaps because they are not familiar with candidates from outside their district. Many others, I suspect, simply don't realize they're allowed to vote for candidates from other districts.
My second suggestion, random ballots, has to do with a rather glaring discrepancy in the outcome of non-partisan and primary elections. In these races, the candidates are listed in alphabetical order. Based on my analysis of recent elections, I estimate that about 10% of the voters in each race simply vote for the first candidate on the ballot. When I ran for the school board three years ago, an additional 10% of the vote wouldn't have been enough for me to win, but if I ever decide to run again in a non-partisan election where the candidates are listed in alphabetical order, I will certainly change my name before filing to something like "Abbot" or even "Aardvark" to make sure I'm first on the ballot.
The problem here is not so much that the candidates are listed in alphabetical order, but that they are listed in the same order on each ballot. The solution is simple-- list the candidates in random order. Thus, if there are four candidates for a primary nomination or non-partisan office, each would be listed first on one fourth of the ballots and voting machines. This might give rise to complaints about the actual distribution of machines on which a particular candidate was listed first, and the total number of votes cast on those machines. It would still be better than the current system. With computerized internet voting, this could be dealt with by randomizing the ballot for each voter, rather than just for each machine. More likely complaints would come from the election commission, which would have to program the machines to correctly tally the randomized ballots. I'm certain this would not be an insurmountable problem, however, and it would be worthwhile because it would clearly make the electon more truly reflective of the public will. I would also anticipate complaints from any group or individual that uses sample ballots for educational or political purposes. Since the candidates on the actual ballots could be in any order, there would be no way to guarantee the sample ballot would be identical to the actual ballot presented to each voter. Those who distribute sample ballots would have to explain that the candidates might be listed in a different order on the actual ballots, so the voter should read the names on the ballot, and not simply vote for candidates that are in the same positions as their choices on the sample ballot.
Again, these changes won't make our system perfect, but I think the cost of implementing them would be more than worthwhile.
John Kozma
1333 South Edgewater Dr.